Artificial intelligence (AI) is a concept that refers to the use of mostly computers and little to no human involvement in task completion. In other words, the goal of AI is to have computers replicate our thought processes in order to increase our productivity in the fast-paced world we live in today.
One of the most significant information technology revolutions has occurred. As far as we can tell, significant advancements have been made in theoretical research and its implementation. The appearance of several robots in various domains, particularly in bioinformatics, is widely acknowledged as being a result of AI.
Several effective robot-assisted surgeries have been performed in conjunction with medical treatment. It improves the accuracy and productivity of medical work. These days, image-based medical diagnosis and screening with AI assistance are becoming more common.
As we all hear, melanoma, a skin cancer could be diagnosed with a computer algorithm based on macro images captured by a common camera. In the field of ophthalmology, especially in the blind-causing diseases, it mainly attributes to medical imaging identification and auxiliary diagnosis.
The application of artificial intelligence in ophthalmology mainly concentrates on the diseases with a high incidence, such as diabetic retinopathy, age-related macular degeneration, glaucoma, retinopathy of prematurity, age-related or congenital cataract and few with retinal vein occlusion.
The two primary areas of AI devices are machine learning techniques and natural language processing techniques. But, up to this point, the former is the auxiliary screening and diagnostic method that we frequently discuss.
By processing the incoming data and generalizing a performance standard, machine learning offers techniques or algorithms that can automatically develop a model of complicated relationships.
And in a nutshell, it enables computers to generate accurate predictions or decisions by continually studying representative existing materials.
Machine learning frequently requires a big amount of training data in order to be able to create an appropriate model. And the most of them require relative competent professionals to preliminarily identify their properties.
Moreover, additional data are used to validate the established method. This indicates that the procedures primarily consist of the training set and validation set. Consequently, gathering a large number of representative training instances is a crucial step.
By processing and analyzing a significant amount of data, artificial intelligence (AI) software may carry out cognitive tasks like learning and problem-solving. In other words, the machine can acquire experience much like people do.
It was founded in 1956 and quickly established itself in a variety of medical specialties, including ophthalmology in the late 1990s when color fundus photography began to acquire prominence in the diagnosis of diabetic retinopathy (DR).
Later on, its use was not limited to but tried extensively in many subspecialties of the eye such as cataract, myopia and glaucoma screening, corneal ectasia, keratoconus, retinopathy of prematurity (ROP) and ocular reconstruction.
It can also be used in calculating intraocular lens power and while planning squint surgery and intravitreal injections. AI can even detect cognitive loss, Alzheimer’s disease and cerebrovascular stroke risk from fundus photographs and optical coherence tomography (OCT).
Millions of people are impacted daily by the illness burden of diabetes mellitus. The current disease load is estimated to be 463 million people, and by 2040, it is expected to increase to 642 million people.
The microvascular problem known as DR affects the blood vessels in the retina, causing irreparable damage and blindness. Regardless of the type of diabetes, these people require an early diagnosis and fast treatment.
The burden on the retina specialists is reduced by routine dilated fundus screening in these patients using ophthalmoscopy and color fundus pictures. The automated grading of DR using ML and DL models, CNN, and the massive-training artificial neural network has demonstrated encouraging results.
ROP is one of the leading causes of childhood blindness throughout the world. This vasoproliferative condition affects preterm infants with low gestational age and those with low birth weight. This condition should be diagnosed promptly so that timely intervention can be done.
This can be abetted with the help of AI, which provides an automated, quantifiable and highly objective diagnosis in plus disease in ROP.
One more area of application of AI in ROP is the utilization of the DL algorithms into medical training to standardize ROP training and education through tele-education. However, there are few clinical and technical challenges in the implementation of AI in the actual scenario.
AMD is considered the leading cause of central vision loss in the elderly age group. The challenges in diagnosing and managing this silent progressive retinal condition have led to the rising prevalence of the disease.
AI has evolved to help in the automated detection of drusens in the very early stages and stratify the disease’s progression. AMD is clinically characterized by the presence of drusens and retinal pigment epithelium changes progressing into geographic atrophy and neovasculari-zation.
High intraocular pressure causes glaucoma, a progressive visual neuropathy that results in the loss of retinal nerve fibers and irreversible blindness. The progression of the disease can be slowed down with early treatment. AI can assist in spotting borderline instances and predicting the progression of the illness.
To identify the disease, ML has been applied in a number of investigations. All the relevant data, such as alterations to the optic disc, intraocular pressure (IOP), gonioscopy, thickness of the retinal nerve fiber layer, visual fields, etc., should be able to be evaluated by a thorough AI for glaucoma.
However, such a comprehensive package is yet to come to the real-time world. The application of AI in measuring IOP is now limited to the Sensimed Triggerfish, a contact lens-based continuous IOP monitoring device that measures the corneal strain changes induced by IOP fluctuations.
Nuclear cataracts can be graded by AI utilizing algorithms based on ML or DL systems that perform as well as a clinician's grading, according to studies on the subject. A system to assess cataracts based on slit-lamp pictures was proposed by Gao et al.
Liu et al. concentrated on accurately and sensitively diagnosing and classifying pediatric cataracts. A universal AI platform and multilevel collaboration pattern were created by Wu et al. to perform well in the diagnostic and referral services for young children and adults with cataracts.
Dong et al. have suggested utilizing a combination of a DL system to extract images (Caffe software), followed by an ML algorithm (known as Softmax function) for severity assessment, to automatically detect and grade cataracts from color fundus photographs. Due to its ability to recognize various cataract surgery phases, AI has also been tested in the training of residents in this field.
With its extensive reliance on imaging, ophthalmology is a leader in the application of artificial intelligence (AI) in healthcare. Although there are many benefits for patients and medical professionals, there are still barriers to completely integrating AI, including concerns about the economy, ethics, and data privacy.
According to Konstantinos Balaskas, MD, FEBO, MRCOphth, a retinal expert at Moorfields Eye Hospital, London, United Kingdom, and director of the Moorfields Ophthalmic Reading Centre and AI Analytics Hub, AI is a broad term.
“The type of AI that has generated a lot of excitement in recent years is called ‘deep learning,’ ” he said. “This is a process by which software programs learn to perform certain tasks by processing large quantities of data.”
Deep learning is what has made ophthalmology a pioneer in the field of implementing AI in medicine, because we are increasingly reliant on imaging tests to monitor our patients.
“Particularly in my subspecialty of interest, medical retina, imaging tests such as optical coherence tomography (OCT) are performed very frequently and have provided the material to train, test, and then apply AI decision support systems,” Balaskas noted.
In retina particularly, some of the most common causes of visual loss in the Western world—such as age-related macular degeneration (AMD) and diabetic retinopathy—require early detection, prompt initiation of treatment, and regular monitoring to preserve vision.
Balaskas said this is where AI decision support systems can help improve access to care and ensure optimal clinical outcomes for patients. Balaskas cited the AI decision support system developed in collaboration between Moorfields Eye Hospital, where he is based, and Google DeepMind.
“It is able to read OCT scans, interpret them, provide a diagnosis, and make management recommendations,” he said. “The other area where AI shows promise is in the development of personalized treatment plans for patients by being able to predict their response to treatment and their visual outcomes over a period of time.”
When considering common conditions that threaten vision, such as AMD and diabetic retinopathy, Balaskas says AI decision support tools—once validated and once they have gained regulatory approval as medical devices—can help improve access to care.
“They can, for example, assist health practitioners in the community in diagnosing diseases early,” he explained.”
“In the United Kingdom, where OCT scans are widely available in high street optician practices, an AI tool would be particularly useful to assist them to interpret scans correctly and identify disease at an early stage.”
Similarly, in diabetic retinopathy, where patients require regular screening and monitoring, AI tools can significantly increase efficiency of screening programs. Balaskas pointed out that such applications already exist and can be of particular use in diabetic retinopathy screening programs such as in underresourced health care settings.
“Other indications for the application of AI monitoring, like AMD, are in advanced stages of development but have not yet been implemented in real life,” he added. Balaskas said there are challenges with integrating AI into retina diagnostics and treatments.
He noted that he has a personal academic interest in implementation science, which looks at the gap between developing a medical device such as an AI decision support tool and deploying it in clinical practice.
“The potential barriers that we need to overcome for the tool to be deployed in a meaningful way to improve outcomes for our patients go beyond testing and validation,” he said. “These include economic evaluations: how would such an automated decision support model affect the finances of a health care system, so that it could provide good value for money or achieve cost savings?”
The next consideration is human factors, particularly how these models of care that rely on AI are perceived and accepted by patients and practitioners. What is the level of trust in these technologies?
What level of information and education of patients and the public is required to build confidence in their use? Then there are considerations regarding training and technical infrastructure to support these tools.
Balaskas notes that ethical and data-privacy issues, as well as medicolegal considerations, are also important. Who is responsible for decisions made by an AI algorithm rather than a human? How do these tools affect the way health care professionals diagnose and treat disease?
“There is a phenomenon called automation bias, where practitioners are sometimes more likely to defer to the recommendation of the AI tool—even perhaps against their better judgement,” he said.
Balaskas notes the issue of interpretability— that in many instances these AI tools are opaque in their functioning. “We do not fully understand how a specific recommendation is reached, whether that is a diagnosis or a management recommendation, and that lack of transparency can exacerbate the medical, legal, and ethical issues that were mentioned earlier,” he pointed out.
“In summary, we have found that there are several hurdles to overcome before AI tools can be deployed in real life in a way that is safe and will improve clinical outcomes.” Moreover, Balaskas notes that life could change for ophthalmologists in the future, and he has a optimistic vision of AI in medical practice.
“Our field is becoming increasingly complex and we need to process data from various sources when we are assessing our patients: data from the many imaging modalities, genetic data and the various types of omics, such as proteomics and the emerging field of oculomics, where features on the eye examination can be indicative of problems with systemic health,” he said.
Balaskas also noted that data from home vision monitoring devices will become increasingly available. However, Balaskas noted that making sense of all this data in order to develop a personalized treatment plan for each patient can be daunting.
“AI could become a very useful aid and, as described in the Topol Review on AI commissioned by Health Education England, provide the gift of time to patients and practitioners, giving them the chance to discuss and decide together what the optimal treatment plan is, informed by the processing of high-dimensional complex data sources,” he concluded.